Automatic Report Generation and Email to Aggregator

[sajitnair]

Hi All,

I configured the auto report generator but it was not working earlier, eventually it started working after I rebooted the usage meter vm. Now it does send the email with the report attached on the choosen day of the month to the aggregator and my email addresses. So far so good, but the report is a Detailed Billing Report with too much details that I dont want to share with the aggregator. There is no option to choose Aggregate Billing Report for automatic reporting. Is there a way to do that? Send Aggregate Billing Report using automatic report generation.

Thanks

SN

[IamTHEvilONE]

You'll have to ask your aggregator exactly which report they require, based on what you have available.

[sajitnair]

Thanks for your response, I believe you did not understand my question. I know what the aggregator wants and that is an Aggregate Billing Report, unfortunately what I am not able to do is use automatic report generation to send the aggregate billing report. There is no option to choose what kind of report should be sent using automatic report generation. It picks the detailed billing report by default and sends it. Hope I am clear this time

[dbriccetti]

Hi. Newer versions of Usage Meter allow you to chose which report(s) to send. Perhaps you are using 2.2?

Dave Briccetti

[sajitnair]

Hi Dave,

Thanks for the tip, 2.3 definitely has the report options we can choose from.

Now though the update was not supported based on the 2.3 user guide, I still went ahead and did a update from 2.2. The colletions ran for couple of vcenters but was was failing for the vcenter which had a larger number of vm's. The error in the appliance log file is below:

ERROR [VM collection] collect.Collector$: java.util.concurrent.ExecutionException: java.lang.OutOfMemoryError: Java heap space

ERROR [VM collection] collect.Collector$: java.util.concurrent.FutureTask$Sync.innerGet(FutureTask.java:222)
java.util.concurrent.FutureTask.get(FutureTask.java:83)

It did not even send an email alert for the failure. The UM virtual machine has 1 vcpu, 2.5 GB vmem. I have created a reservation and it has no limits.

I then went ahead and created an absolutely new vm and installed 2.3.2 and it worked for 1 sample, but since then it gives the same out of memory error. I understand that this could be due to lack of memory and tried 3 gb instead and it still fails.

Thanks,

Sajit

[dbriccetti]

Curious! We have tested with tens of thousands of VMs with the current memory settings.

Could you do an experiment, please? Find this file

/opt/vmware/cloudusagemetering/scripts/tomcat

/etc/init.d/tomcat

and increase the value here: -Xmx1536m

Then restart Tomcat

service stop tomcat

Tomcat may not stop properly. To work around this you can:

• Restart the Usage Meter VM, or
• Use this command: top -u usgmtr, and kill any processes you see

service start tomcat

Message was edited by: dbriccetti. Corrected file name.

[sajitnair]

I changed it to  -Xmx2536m and rebooted the UM, but no luck it still fails on the same error. Also did recycle tomcat as you suggested before the reboot, guess I will have to revert to 2.2 now, it was collecting, just not reporting the way I wanted it to.

[dbriccetti]

Another thing you can do to lower memory requirements is serialize your collections. Edit /usr/local/apache-tomcat-7.0.14/webapp/WEB-INF/web.xml and change the 5 to a 1.

<context-param>
        <param-name>maxConcurrentCollections</param-name>
        <param-value>5</param-value>
</context-param>

Restart Tomcat or the VM.

You can also go higher than -Xmx2536m if you like.

If this persists, perhaps you can send me your /var/log/usgmtr/um.log around where the error occurs.

Or explain what you mean by upgrading from 2.2 even though the user guide said you can’t. We have a procedure for upgrading from 2.2.

Thanks for helping us get to the bottom of this.

Dave “Bric” Briccetti

[e-mail]

[sajitnair]

So, here is what I have done. Installed a new 2.2 UM and it is collecting (hosted on the same ESXi, same datastore and with 1 vcpu and 2.5 gb vmem). I have left the 2.3 running for now to test and further diagnostics. Let me know if you have anything else that I can try. I would really love the automated reporting :smileyhappy: